Kidney International, Vol. 55 (1999), pp. 160–167

Renal proximal tubular cell fibronectin accumulation inresponse to glucose is polyol pathway dependent

KIMBERLEY MORRISEY, ROBERT STEADMAN, JOHN D. WILLIAMS, and ALED O. PHILLIPS

Institute of Nephrology, University of Wales College of Medicine, Cardiff Royal Infirmary, Newport Road, Cardiff, Wales,United Kingdom

Renal proximal tubular cell fibronectin accumulation in response to velop in the absence of measurable dysfunction [1]. Oneglucose is polyol pathway dependent. of the earliest renal pathological changes in diabetes isBackground. Thickening and reduplication of the tubular basement

an increase in tubular basement membrane mass thatmembrane have been reported as early events in diabetic nephropathy.In this study, we have examined the polar requirements of proximal accompanies the development of renal hypertrophy [2].tubular cells for the d-glucose–stimulated accumulation of fibronectin These changes are strongly correlated with glycemic con-and the mechanism by which this occurred, with particular emphasis

trol and occur independently of the pathological changeson the polyol pathway.Methods. To determine the polarity of fibronectin generation in associated with progressive diabetic nephropathy [3]. It

response to glucose, LLC-PK1 cells were grown on porous tissue culturesuggests that alterations in renal structure may occurinserts. Monolayer confluence was determined by serial measurement

of transepithelial resistance. Confluent cells were growth arrested by that are not specific to nephropathy but reflect a conse-serum deprivation, and all experiments were performed under serum- quence of long-standing diabetes/hyperglycemia. Wefree conditions.

and others have demonstrated that in vitro exposure ofResults. Application of 25 mm d-glucose to either the apical orbasolateral aspect of LLC-PK1 cells led to fibronectin accumulation in human renal proximal tubular cells to 25 mm d-glucosethe basolateral compartment. This reached statistical significance 24 increased generation of type IV collagen and fibronectinhours following apical addition of glucose (2.6-fold increase compared

[4–6]. Our observations suggested that the accumulationwith 5 mm d-glucose, P 5 0.0025, N 5 6 vs. N 5 4 controls) and12 hours after the basolateral addition of glucose (2.5-fold increase of these basement membrane constituents was relatedcompared with 5 mm d-glucose, P 5 0.03, N 5 6 vs. N 5 4 controls).

to alterations in matrix degradation [5].Exposure of cells to glucose at either their apical or basolateral aspectleads to accumulation of intracellular glucose and polyol pathway acti- In diabetes, proximal tubular cells may be exposed tovation, as assessed by sorbitol accumulation. The increase in fibronectin conditions of elevated glucose either apically as a resultconcentration in response to glucose was inhibited by the aldose reduc-

of glycosuria or basely as a result of elevated interstitialtase inhibitor sorbinil. At a dose of 100 mm sorbinil, there was a 59%inhibition of fibronectin accumulation in response to apical glucose (P 5 tissue concentrations of glucose. Our studies on the mech-0.004, N 5 3 sorbinil vs. N 5 4 controls) and a 66% inhibition in anism of glucose induction of transforming growth factorresponse to basolateral glucose (P 5 0.008, N 5 3 sorbinil vs. N 5 4controls) 48 hours after the addition of the inhibitor. Furthermore, b1 (TGF-b1) in proximal tubular cells demonstrated thatfibronectin accumulation was also demonstrated following both the induction of TGF-b1 mRNA was a polar phenomenon,apical and basolateral addition of 1 mm sorbitol, but not following the

occurring only following the addition of glucose to theaddition of 25 mm galactose to either aspect of the cells. Followingthe addition of sorbitol, there was a 2.8-fold increase in fibronectin 48 basolateral aspect of the cells [7].hours after apical stimulation (P 5 0.01, N 5 3 treated vs. N 5 4 In this study, we have examined the polar requirementscontrol) and a 2.27-fold increase following basolateral stimulation (P 50.04, N 5 3 treated vs. N 5 4 control) at 24 hours. of proximal tubular cells for the d-glucose–stimulated

Conclusions. In summary, these data demonstrate that fibronectin accumulation of fibronectin using the porcine LLC-PK1generation in response to glucose was nonpolar in terms of applicationtubular cell line, grown in polarized culture on tissueof glucose but was polar in terms of fibronectin accumulation. The

mechanisms of glucose-induced modulation of fibronectin were medi- culture inserts. We also examined the mechanism byated by polyol pathway activation and were more specifically related which glucose led to the accumulation of fibronectin into the metabolism of sorbitol to fructose.

this polarized system, with particular emphasis on thepolyol pathway.

Diabetic nephropathy is characterized by glomerular,vascular, tubular, and interstitial lesions that initially de- METHODS

d- and l-isoforms of glucose, sorbitol, galactose, andReceived for publication April 13, 1998

a-methyl d-glucose were purchased from Sigma Chemi-and in revised form August 4, 1998Accepted for publication August 12, 1998 cals (Poole, UK). 2-Deoxy d-glucose was obtained from

Fluka Chemicals (Gillingham, UK). The aldose reduc- 1999 by the International Society of Nephrology

160

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation 161

tase inhibitor sorbinil was a gift from Pfizer Central Re- RESULTSsearch (Pfizer Inc., Groton, CT, USA). Polarity of D-glucose–stimulated fibronectin

Supernatant glucose and sorbitol concentrations and generationintracellular glucose, fructose, and sorbitol concentra- Confluent monolayers of growth-arrested LLC-PK1tions were determined by commercial assays (Boeh- cells, on porous tissue culture inserts, were exposed toringer Mannheim, Lewes, UK). For determination of 5 or 25 mm d-glucose or 25 mm l-glucose on either theirintracellular glucose, fructose, and sorbitol, cells were apical aspect or 25 mm a-methyl glucose (apically only)lyzed by the addition of ice-cold sterile water. Prior to or 25 mm 2-deoxy d-glucose (basolaterally only) for upanalysis, total cell protein was determined by BioRad to 48 hours.protein assay (Bio-Rad Laboratories, Hemel Hemp- Increased fibronectin concentration was detected instead, UK). Samples were subsequently centrifuged for the basolateral compartment following the addition of10 minutes at 4,000 r.p.m. to remove cell debris. Superna- 25 mm d-glucose to either the apical or basolateral aspecttant was removed, and glucose analysis was performed. of the cells. This reached statistical significance 24 hours

Fibronectin concentration in culture supernatant was following the addition of apical glucose (2.6-fold increasedetermined using an antibody competition ELISA as compared with 5 mm d-glucose, P 5 0.0025, N 5 6 vs.previously described [8]. In all experiments, there was N 5 4 controls), and 12 hours after the addition of baso-no difference in either the total cell protein or total cell lateral glucose (2.54-fold increase compared with 5 mmnumber in confluent cell monolayers grown on tissue d-glucose, P 5 0.03, N 5 6 vs. N 5 4 controls; Fig. 1).culture inserts. In contrast, there was no increase in the concentration

of fibronectin in the apical compartment following theCell cultureaddition of 25 mm d-glucose to either the apical or the

Experiments to study the polarity of stimulation and basolateral aspects of the cells. Furthermore, there wassecretion of fibronectin were performed using LLC-PK1 no increase in the concentration of fibronectin in eithercells, which are a widely used porcine cell line of proximal compartment following the addition of 25 mm l-glucose,tubular origin. Growth medium for cell culture com- or either of the transported, nonmetabolized glucoseprised M199 with 10% fetal calf serum (Life Technolo- analogues a-methyl glucose or 2 deoxy d-glucose (Fig. 1).gies Ltd, Paisley, Scotland, UK). Cells were grown at378C in 5% CO2 and 95% air. For the determination of Effect of glucose on polyol pathway activationpolarity of the effects of d-glucose, cells were seeded onto In other cell systems, it has been demonstrated thatporous tissue culture inserts (0.4 mm pore size; Becton elevated glucose may alter the turnover of extracellularDickinson, Oxford, UK). Reassembly of tight junctions matrix components by activation of the polyol pathwayand integrity of the monolayer were demonstrated by (Fig. 2). Despite the low affinity of aldose reductase forserial measurements of transepithelial resistance using a glucose, an increase in intracellular glucose concentra-commercially available system (Milliceld-ERS Resistance tion will lead to an increase in flux of glucose throughSystem; Millipore, Bedford, MA, USA) as previously the polyol pathway. Initial experiments were thereforedescribed [9]. Confirmation that peak increases in trans- performed to examine the effect of the addition of ele-epithelial resistance were reliable indicators of the for- vated d-glucose in our polarized system on intracellularmation of a confluent monolayer was demonstrated by glucose and sorbitol concentrations.the correlation of transepithelial resistance measurements Figure 3 demonstrates the effect of the addition of 25to diffusion of both phenol red and 125I-labeled albumin mm d-glucose to the apical aspect of the cells. The glucoseacross the monolayer as previously described [7]. All concentration in the apical compartment fell steadilyexperiments were performed on confluent monolayers of with time as glucose was transported from the apicalcells, under serum-free conditions, following a “growth compartment to the basolateral compartment. There wasarrest” period of 72 hours in serum-free medium. No a corresponding increase in the concentration of glucosechanges in transepithelial resistance were demonstrated in the basolateral compartment. In conjunction with theover the experimental period. increase in glucose flux through the cell, there was also

an increase in intracellular glucose concentration thatStatistics was apparent as early as one hour following addition of

For each individual experiment, the mean of duplicate 25 mm d-glucose to the apical aspect of the cell. Forty-determinations was calculated. The data are presented eight hours after the addition of 25 mm d-glucose to theas means 6 sd of N experiments. Statistical analysis apical aspect of the cells, there was a 3.06-fold increasebetween two groups was performed using the unpaired in intracellular glucose concentration (P 5 0.0001, N 5 3).two-tailed Student’s t-test, with a value of P less than Figure 4 illustrates the effect of addition of 25 mm

d-glucose to the basolateral aspect of the cell. There was0.05 considered to represent a significant difference.

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation162

a steady decline in the concentration of glucose in thebasolateral compartment, but there was no correspond-ing increase in apical glucose concentration over the timecourse of the experiments. Again, there was an increasein intracellular glucose following the addition of basolat-eral d-glucose that was comparable to the increase inintracellular glucose seen following the addition of glu-cose to the apical aspect of the cells. Forty-eight hoursafter the addition of 25 mm d-glucose to the basolateralaspect of the cells, there was a 2.8-fold increase in intra-cellular glucose concentration (P 5 0.0006, N 5 3).

To demonstrate activation of polyol pathway follow-ing the addition of glucose, confluent growth arrestedcells were exposed to 25 mm d-glucose for up to 48 hours.At each time point, cells were lyzed by the addition ofice-cold water and intracellular sorbitol measured. Un-der these conditions, there was a time-dependent in-crease in intracellular sorbitol, which became significant48 hours after the addition of 25 mm d-glucose (P 50.02, N 5 3; Fig. 5). Furthermore, following the additionof 25 mm d-glucose to either the apical or basolateralaspect of the cells, there was also a time-dependent in-crease in intracellular fructose concentrations (Fig. 6).

Effect of polyol pathway inhibition

The effect of inhibition of polyol pathway on glucose-stimulated fibronectin accumulation was examined bythe addition of 5 or 25 mm d-glucose to confluent growth-arrested cells in the presence of increasing doses of thealdose reductase inhibitor sorbinil. Initial experimentsto confirm the inhibitory effect of the inhibitor on polyolpathway activity were performed by the addition of 25mm d-glucose together with sorbinil to confluent growth-arrested cells grown in conventional two-dimensionalculture conditions. These experiments demonstrated adose-dependent decrease in intracellular sorbitol con-centration, confirming inhibition of polyol pathway activ-ity (Fig. 7A). In our polarized culture system, there wasalso a dose-dependent abrogation of 25 mm d-glucose–induced fibronectin accumulation concentration gener-ated following the addition of 25 mm d-glucose whensorbinil was added simultaneously to the apical and baso-lateral compartments (Fig. 7B). At a dose of 100 mmsorbinil, there was 59% inhibition of fibronectin accumu-lation in response to apical glucose (P 5 0.004, N 5 3

Fig. 1. Polarity of stimulation of fibronectin. Confluent growth arrestedsorbinil vs. N 5 4 control) and a 66% inhibition in re-cell monolayers were stimulated on either the apical (A) aspect or their

basolateral (B) aspect of for up to 48 hours, under serum free conditions sponse to basolateral glucose (P 5 0.008, N 5 3 sorbinilwith 5 mm l-glucose ( ), 25 mm d-glucose ( ), 25 mm a-methyl glucose vs. N 5 4 control) 48 hours after the addition of the( ), (apical only), 25 mm 2-deoxy-glucose ( ), (basolateral only) or

inhibitor.25 mm d-glucose ( ). Medium on the “unstimulated” side of the cellscontained 5 mm d-glucose. Supernatant samples were collected andfibronectin concentration in the apical compartment and basolateral Mechanism of induction of fibronectincompartment determined by ELISA. No increase in fibronectin was

Biological consequences of polyol pathway activationdemonstrated in the apical compartment, and results shown representanalysis of supernatant samples taken from the basolateral aspect of may be the result of either polyol accumulation or thethe cells. Results represent mean 6 sd (*P , 0.05; N 5 6 experimental consequence of conversion of sorbitol to fructose. Thevs. N 5 4 5 mm d-glucose controls).

mechanism by which glucose led to an increased genera-

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation 163

Fig. 2. The polyol/aldose reductase pathway.

Fig. 3. Intracellular glucose accumulation following apical exposure toglucose. Following a growth arrest period of 72 hours, 25 mm d-glucose

Fig. 4. Intracellular glucose accumulation following basolateral expo-was added to the apical compartment of confluent cells, with freshsure to glucose. Following a growth arrest period of 72 hours, 25 mmmedium containing 5 mm d-glucose added to the opposite compartment.d-glucose was added to the apical compartment of confluent cells, withSupernatant samples from apical ( ) and basolateral ( ) compart-fresh medium containing 5 mm d-glucose added to the opposite compart-ments were collected over the next 48 hours for determination of glucosement. Supernatant samples from apical ( ) and basolateral ( ) com-concentrations. At each time point cells were lyzed by the addition ofpartments were collected over the next 48 hours for the determinationice cold sterile water prior to the determination of intracellular glucoseof glucose concentrations. At each time point cells were lyzed by theconcentration (h). Results represent mean 6 sd of 3 separate experi-addition of ice-cold sterile water prior to the determination of intracellu-ments (*P , 0.05).lar glucose concentration (h). Results represent the mean 6 sd of 3separate experiments, *P , 0.05.

tion of fibronectin was examined by the addition of 5 mmfibronectin concentration in either apical or basolaterald-glucose, 25 mm galactose, or 1 mm sorbitol to eithercompartments following the addition of 25 mm galactosethe apical or basolateral aspects of confluent LLC-PK1to the apical compartment.cells grown on tissue culture inserts under serum-free

Similarly, following application of either 1 mm sorbitolconditions.to the basolateral aspect of the cells, there was a time-Following the addition of sorbitol to either the apicaldependent increase in the fibronectin concentration inor basolateral aspect of the cells, there was a time-depen-the supernatant taken from the basolateral aspect (Fig.dent increase in intracellular sorbitol concentration (Fig.9B). Twelve hours after the addition of sorbitol, this8). Increased concentration of fibronectin in the superna-represented a 3.1-fold increase in fibronectin concentra-tant taken from the basolateral compartment was seention (P 5 0.01, N 5 3 vs. N 5 4 5 mm d-glucose controls).following the addition of 1 mm sorbitol to either aspectThere was no increase in fibronectin concentration inof the cells (Fig. 9A). Following apical stimulation, thisthe basolateral compartment following the addition of

was significant 48 hours after the addition of sorbitol 25 mm galactose basolaterally.and represented a 2.8-fold increase in fibronectin concen-tration (P 5 0.001, N 5 3 vs. N 5 4 5 mm d-glucose

DISCUSSIONcontrols). No increase in fibronectin was seen in theapical compartment following apical addition of either Alterations in tubular basement membrane (TBM)

composition occur early in the course of diabetic ne-5 mm d-glucose or 1 mm sorbitol. There was no increase in

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation164

Fig. 5. Polyol pathway activation by 25 mM D-glucose. Twenty-five mmd-glucose ( ) or 25 mm l-glucose (h) was added to confluent growtharrested cells under serum free medium. At each time point followingthe removal of supernatant, and washing of the cell monolayer withPBS, cells were lyzed by addition of ice cold water, and intracellularsorbitol measured. Results represent mean 6 sd of 3 separate experi-ments (*P , 0.05).

Fig. 7. Polyol pathway inhibition. (A) Confluent growth arrested cellswere stimulated with 25 mm d-glucose in the presence of increasingdoses of the aldose reductase inhibitor sorbinil. At each time pointfollowing removal of supernatant, and washing of the cell monolayerwith PBS, cells were lyzed by addition of ice cold water, and intracellularsorbitol measured. Results represent mean 6 sd, N 5 3 sorbinil vs. N 5 4control (*P , 0.05). (B) Confluent growth arrested cells were stimulatedwith 25 mm d-glucose either apically (h) or basolaterally (j) with theFig. 6. Intracellular fructose accumulation following exposure to 25aldose reductase inhibitor sorbinil, added to both compartments. CellmM D-glucose. Following a growth arrest period of 72 hours, 25 mmculture supernatant was collected after 48 hours and fibronectin concen-d-glucose was added to either the apical (j) or basolateral (h) compart-tration in the basolateral compartment quantitated by ELISA. Resultsment of confluent cells. At each time point, cells were lyzed by therepresent mean 6 sd, N 5 3 sorbinil vs. N 5 4 control (*P , 0.05).addition of ice-cold sterile water prior to the determination of intracellu-

lar fructose concentration. Results represent mean 6 sd of 3 separateexperiments (*P , 0.05).

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation 165

Fig. 8. Intracellular sorbitol accumulation following exposure to 1 mM

sorbitol. Following a growth arrest period of 72 hours, 1 mm sorbitolwas added to the either the apical (j) or basolateral (h) compartmentof confluent cells. At each time point cells were lyzed by the addition ofice-cold sterile water prior to the determination of intracellular sorbitolconcentration. Results represent mean 6 sd of 3 separate experiments(*P , 0.05).

phropathy, and recent reports suggest that these changesmay precede any histologic alterations within the glo-merulus, as assessed by light microscopy [2]. Both invitro and in vivo studies have implicated elevation ofglucose concentration in the increased production ofbasement membrane components by endothelial [10–12]and mesangial cells [13–15]. We have previously demon-strated that exposure of renal proximal tubular cells toelevated concentrations of glucose led to the accumula-tion of fibronectin and type IV collagen by a mechanismthat involved an alteration in the degradative pathwayfor these matrix constituents. These studies demon-strated no effect of glucose on type IV collagen or fibro-nectin gene expression, although increased synthesis bya post-transcriptional mechanism may contribute to the

Fig. 9. Mechanism of fibronectin accumulation. Confluent growth ar-increase in concentration of fibronectin and collagenrested cell monolayers were stimulated on either the apical (A) aspectseen in the tissue culture supernatant following the addi-or their basolateral (B) aspect of for up to 48 hours, under serum free

tion of 25 mm d-glucose. conditions by the addition of either 5 mm d-glucose (h), 25 mm Galac-tose ( ) or 1 mm sorbitol (j). Medium on the “unstimulated” side ofIn vivo, proximal tubular cells have an asymmetricthe cells contained 5 mm d-glucose. The medium in the “stimulated”distribution of transport proteins that facilitates the reab-compartment also contained 5 mm d-glucose. Results demonstrate fi-

sorption of water, glucose, and electrolytes from the glo- bronectin concentration in the supernatant collected from the basolat-eral compartment, and represent mean 6 sd, N 5 3 vs. N 5 4 5 mmmerular filtrate. We and others have demonstrated thatd-glucose controls (*P , 0.05).an alteration of proximal tubular cell function may occur

in a polarized fashion [7, 16]. In this article, we haveexamined the polarity of stimulation by glucose as wellas the polarity of secretion of fibronectin. These datademonstrate that exposure of cells at either apical or baso-

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation166

lateral surface to elevated glucose concentrations led to for fibronectin accumulation, as metabolism of galactoseto its polyol galactiol would also lead to a decrease inextracellular accumulation of fibronectin and that this

increase was seen in supernatant taken from only the the NADPH/NADP ratio. The likely mechanism, there-fore, is related to the metabolism of sorbitol to fructose.basolateral aspect of the cells. The mechanism by which

glucose led to the accumulation of fibronectin was unre- Increased metabolism of sorbitol is accompanied by anincrease in the NADH/NAD ratio. This may impact onlated to alterations in osmolarity, as these effects were

not reproduced following the addition of l-glucose (the the activity of numerous cytoplasmic and mitochondrialenzymes that use NADH and NAD as cofactors and,stereoisomer of glucose that is not transported into the

cell). Furthermore, using a-methyl glucose and 2-deoxy- hence, modulate the activity of many metabolic path-ways. One consequence of an alteration of NADH/NADglucose as glucose analogues, which are transported

across the apical and basolateral aspects of the cell, re- ratio is an increase in de novo diacylglycerol generationand associated activation of protein kinase C. This isspectively, but are not metabolized beyond their phos-

phorylation, we have demonstrated that both transport unlikely to explain the effects described in this manu-script, as we have previously demonstrated that fibro-and metabolism of glucose are required for the accumu-

lation of fibronectin. nectin accumulation in response to glucose was not in-fluenced by inhibition of protein kinase C [5]. OtherA common mechanism by which glucose modulates

cell function is by an increased polyol pathway activity possible consequences of an increase in the cytosolicNADH/NAD ratio that could be involved include inhibi-[4, 17, 18]. Metabolism of glucose by the polyol pathway

consists of two reactions. The first reduction of glucose tion of fatty acid oxidation, increased prostaglandin syn-thesis, increased free radical production, and increasedto sorbitol uses nicotinamide adenine dinucleotide phos-

phate (NADPH) as the hydrogen donor and leads to a nitric oxide production. The final consequence of in-creased flux through the polyol pathway that could ac-decrease in the NADPH/NADP ratio. In the second

reaction, sorbitol is oxidized to fructose. This uses NAD count for glucose-mediated alterations in matrix me-tabolism may be the increased formation of fructose.as the hydrogen acceptor and leads to an increase in the

NADH/NAD ratio. Aldose reductase has a low affinity Nonenzymatic fructosylation of tissue occurs at a higherrate when compared with glycosylation, and accumula-for glucose, and at normal glucose concentrations, me-

tabolism of glucose by this pathway represents only a tion of fructose may provide a link between the aldosereductase pathway and “nonenzymatic glycation.”small proportion of glucose metabolism. However, under

conditions of elevated ambient glucose concentrations, The early clinical manifestations of renal disease indiabetes mellitus are likely to reflect a consequence ofintracellular glucose concentration rises, as we have dem-

onstrated, and hence glucose metabolism by the polyol long-standing diabetes. These changes of mild increasein mesangial volume fraction that accompany glomerularpathway increases, leading to the observed increase in

intracellular sorbitol and fructose. Polyol pathway involve- hypertrophy and basement membrane thickening maybe found in patients with normal urinary albumin andment in fibronectin generation was confirmed in this

study by the demonstration that inhibition of the polyol can be divorced from the pathological changes associatedwith progressive diabetic nephropathy. The data in thispathway using the aldose reductase inhibitor sorbinil

led to a reduction in fibronectin concentration in the article describe potential mechanisms that may be in-volved in the changes associated with hyperglycemia,supernatant taken from cells exposed to 25 mm d-glucose.

Numerous mechanisms have been proposed by which which is in contrast with our previous studies examiningglucose modulation of proximal tubule cell (PTC) cyto-increased polyol pathway activity may contribute to the

complications of diabetes [19–22]. The first implicates kine generation [23]. In these studies, induction of TGF-b1mRNA occurred only after application of glucose to theosmotic changes related to the accumulation of sorbitol.

Although we have demonstrated accumulation of intra- basolateral aspect of the cells. In addition, induction ofTGF-b1 secretion required sequential stimulation ofcellular sorbitol following addition of glucose, this is

unlikely to account for the changes seen, as they were not cells with glucose and a second stimulus such as platelet-derived growth factor, with both stimuli applied to thereproduced by the addition of galactose. Unlike sorbitol

(which is readily oxidized to fructose), the polyol of basolateral aspect of the cell [7]. Secretion of TGF-b1in these experiments occurred to both apical and basolat-galactose, galactiol, is not metabolized further, except in

the liver. Therefore, an addition of galactose will lead eral aspects of the cells. These observations, demonstra-ting a lack of TGF-b production following the additionto accumulation of polyol, but from our data, this did

not lead to the accumulation of fibronectin. Similarly, the of 25 mm d-glucose, are therefore consistent with studiesin primary cultures of human proximal tubular cells,addition of galactose did not affect intracellular fructose

concentration (data not shown). Another postulated me- which demonstrated no inhibition of type IV collagenor fibronectin generation in response to 25 mm d-glucosediator of polyol pathway effect is NADPH utilization.

This is also unlikely to be the mechanism responsible following the addition of blocking antibodies to TGF-b

Morrisey et al: Polyol pathway mediates tubular cell fibronectin generation 167

Exposure of human renal proximal tubular cells to glucose leads[5]. The mechanisms of induction of TGF-b1 mRNAto accumulation of type IV collagen and fibronectin by decreased

and protein synthesis and that of modulation of matrix degradation. Kidney Int 52:973–984, 1997turnover by glucose are therefore completely different. 6. Ziyadeh FN, Sharma K, Ericksen M, Wolf G: Stimulation of

collagen gene expression and protein synthesis in murine mesangialIt is interesting to speculate that these differences maycells by high glucose is mediated by autocrine activation of trans-reflect differences in the metabolism of glucose. These forming growth factor beta. J Clin Invest 93:536–542, 1994

data presented in Figures 3 and 4 demonstrate similar 7. Phillips AO, Morrisey R, Steadman R, Williams JD: Polarityof stimulation and secretion of TGF-b1 by proximal tubular cells.accumulation of intracellular glucose following apical orAm J Pathol 150:1101–1111, 1997basolateral application of glucose. This is consistent with 8. Engval E: Fibronectin. Weinhiem, Verlag Chemie, 1986, pp

subsequent activation of polyol pathway under both of 201–2119. Jovov B, Wills NK, Lewis AA: A spectroscopic method for as-these conditions. Although intracellular glucose concen-

sessing confluence of epithelial cell cultures. Am J Physiol 262:trations are similar, it is clear that the metabolism of C1196–C1203, 1991glucose by the cells is greater when glucose is applied 10. Cagliero E, Maiello M, Boeri D, Roy S, Lorenzi M: Increased

expression of basement components in human endothelial cellsto their basolateral aspect. This difference in metabolismcultured in high glucose. J Clin Invest 82:735–738, 1988cannot be explained by metabolism of glucose by the

11. Cagliero E, Roth T, Maiello M, Lorenzi M: Expression of genespolyol pathway. Other metabolic events that may modu- related to the extracellular matrix in human endothelial cells. J

Biol Chem 266:14244–14250, 1991late polarized modulation of cell function include the12. Cagliero E, Roth T, Roy S, Lorenzi M: Characteristics and mech-hexosamine pathway [24] or increased metabolism of

anisms of high-glucose induced over expression of basement mem-glucose by glycolysis [25–27]. brane components in cultured human endothelial cells. Diabetes

40:102–110, 1991In summary, these data in our study illustrate mecha-13. Danne T, Spiro MJ, Spiro RG: Effect of high glucose on type IVnisms by which glucose may be involved in the changes

collagen production by cultured glomerular epithelial, endothelial,of long-standing diabetes. Specifically, we have demon- and mesangial cells. Diabetes 42:170–177, 1993strated that fibronectin generation in response to glucose 14. Roy S, R S, Cagliero E, Lorenzi M: Over expression of fibronectin

induced by diabetes or high glucose: Phenomenon with a memory.was nonpolar in terms of application of glucose, butProc Natl Acad Sci USA 87:404–408, 1990polar in terms of fibronectin accumulation. Furthermore, 15. Nahman NS, Leonharat KL, Cosio FG, Hebert CL: Effects of

modulation of fibronectin was mediated by polyol path- high glucose on cellular proliferation and fibronectin productionby cultured human mesangial cells. Kidney Int 41:396–402, 1992way metabolism of glucose, and more specifically related

16. Burton CJ, Combe C, Walls J, Harris KPG: Fibronectin produc-to the metabolism of sorbitol to fructose. These observa-tion by human tubular cells: The effect of apical protein. Kidney

tions, together with our previous published data, demon- Int 50:760–767, 199617. Beyer-Mears A, Ku L, Cohen MP: Glomerular polyol accumula-strate the important role that proximal tubular cells may

tion on diabetes and its prevention by oral sorbinil. Diabetesplay in the renal histopathological changes associated33:604–607, 1984

with diabetes mellitus. 18. Mower P, Aynedjiam H, Silverman S, Wilkes B, Bank N: Sorbinilprevents glomerular hyper-perfusion in diabetic rats. (abstract)Kidney Int 35:433, 1989ACKNOWLEDGMENTS

19. Larkins RG, Dunlop ME: The link between hyperglycaemia andK.M. is supported by the British Diabetic Association. R.S. is sup- diabetic nephropathy. Diabetologia 35:499–504, 1992

ported by the Kidney Research Unit for Wales Foundation, and A.O.P. 20. Williamson JR, Chang K, Frangos M, Hasan KS, Ido Y, Kawa-is in receipt of a Wellcome Advanced Training Fellowship. The authors mura T, Nyengaard JR, Enden M, Kilo C, Tilton RG: Hypergly-wish to thank Dr. N. Topley and Dr. M. Davies for their support and caemia pseudohypoxia and diabetic complications. Diabetes 42:advice. 801–813, 1993

21. Ruderman NB, Williamson JR, Brownlee M: Glucose and dia-Reprint requests to: Dr. Aled O. Phillips, Institute of Nephrology, betic vascular disease. FASEB J 6:2905–2914, 1992

Cardiff Royal Infirmary, Newport Road, Cardiff CF2 1SZ, Wales, 22. Tilton RG, Baier LD, Harlow JE, Smith SR, Ostrow E, Wil-United Kingdom. liamson JR: Diabetes-induced glomerular dysfunction: Links to aE-mail: PhillipsAO@cf.ac.uk more reduced cytosolic ratio of NADH/NAD. Kidney Int 41:778–

788, 199223. Phillips AO, Steadman R, Topley N, Williams JD: ElevatedREFERENCES d-glucose concentrations modulate TGF-b1 synthesis by human

cultured renal proximal tubular cells: The permissive role of plate-1. Mauer SM, Steffes MW, Ellis EE, Sutherland DER, Brownlet derived growth factor. Am J Pathol 147:362–374, 1995DM, Goetz FC: Structural-functional relationships in diabetic ne-

24. Kolm-Litty V, Sauer U, Nerlich A, Lehmann R, Schlericherphropathy. J Clin Invest 74:1143–1155, 1984ED: High glucose induced transforming growth factor b1 produc-2. Ihm CG, Lee GSL, Nast CC, Artishevsky A, Guillermo R, Levintion is mediated by the hexosamine pathway in porcine glomerularPS, Glassock RJ, Adler SG: Early increased renal procollagenmesangial cells. J Clin Invest 101:160–169, 1998a1 (IV) mRNA levels in streptozotocin induced diabetes. Kidney

25. Craven PA, DeRubertis FR: Protein kinase is activated in glomer-Int 41:768–777, 1992uli from streptozotocin diabetic rats. J Clin Invest 83:1667–1675,3. Brito PL, Fioretto P, Drummond K, Kim Y, Steffes MW, Basgen1989JM, Sisson-Ross S, Mauer M: Proximal tubular basement mem-

26. Wolf BA, Williamson JR, Easom RA, Chang K, Sherman WR,brane width in insulin-dependent diabetes mellitus. Kidney IntTurk J: Diacylglycerol accumulation and microvascular abnormali-53:754–761, 1998ties induced by elevated glucose levels. J Clin Invest 87:31–38, 19914. Bleyer AJ, Fumo P, Snipes ER, Goldfarb S, Simmons DA, Ziya-

27. Brand KA, Hermfisse U: Aerobic glycolysis by proliferating cells:deh FN: Polyol pathway mediates high glucose induced collagensynthesis in proximal tubule. Kidney Int 45:659–666, 1994 A protective strategy against reactive oxygen species. FASEB J

5. Phillips AO, Morrisey K, Martin J, Williams JD, Steadman R: 11:388–395, 1997